基于BP神经网络的谐波电流建模及预估模型

Modeling and Prediction Model of Harmonic Current Based on BP Neural Network

伴随着大数据、人工智能等信息系统的集成与应用,大型建筑物中的能源需求预测、多能源调度、电能质量控制等逐渐成为保障智慧建筑各系统稳定运行的关键技术。电能质量控制方面主要考虑谐波分析与治理,在对大量实测数据进行统计与分析的基础上,针对谐波电流,利用非线性回归与BP神经网络进行建模,分别对电流总谐波畸变率与K系数,电流总谐波畸变率与变压器容量、负载率等关系进行拟合与预估,取得较好结果。本研究对于进一步分析并提升电能质量有重要借鉴作用。

With the integration and application of information systems such as big data and artificial intelligence,energy demand prediction,scheduling of multiple energy sources,and power quality control in large buildings gradually become key technologies to ensure the stable operation of various systems in smart buildings.Harmonic analysis and control are the main considerations in terms of power quality control.Based on the statistics and analysis of a large number of measured data,this article uses nonlinear regression and BP neural network to model the harmonic current and respectively fit and predict the relationship between the total harmonic distortion of current and K coefficient,and between total harmonic distortion of current and transformer capacity,load factor,etc.,with good effect.This study has important reference value for further analysis and improvement of power quality.